calinb

I hope there's enough traffic in this forum group to obtain some advice here. The Ultimaker.com forum seems like the best fit of all the forums I've joined for this question. My daughter just joined a 4H robotics project group. I could spend a lot of money on experimenter or development kits, but I'm wondering if there's any kind of virtual software environment that would run on her Linux OS laptop. I can envision an educational tool with virtual machines (robots?) living in a graphical world that would respond to code that a student would create to program them to do various things. Any ideas here would be much appreciated! -Cal

I'm very pleased to hear that the connector fix solved your problem, carinac! That is where I find the root cause of your printer's symptoms about 80% of the time. The other 20% of the time, culprit is the thermocouple connection at the two little screws in the green connector block, which is the first place I always check, nonetheless, because it's much easier than pulling the three-wire white connector block apart to inspect and crimp on new metal pins. Even when the wires appear to be attached to the pins, sometimes it's only the insulation that's attached at the insulation crimp point and the electrical crimp point or wire is still an intermittent open circuit and unreliable. (There are two crimp points on each pin--one for just the stripped metal multi-strand wire core and one for the wire WITH the insulation covering it.) I always give each wire a good pull from each pin during an inspection. If the entire wire slips out of the pin, it was not a reliably tight connection at the pin and needed cutting, re-stripping, and replacing anyway. If the wire (the plastic insulation) stretches and breaks at or above the insulation strain relief point, the wire was broken inside the insulation and the wire needs to be re-stripped and crimped to a new pin anyway. I've actually experienced a time or two where I thought I'd found the culprit wire and repaired it, only to find out that ANOTHER wire or connection was also defective! Now I just pull on all three wires, one at a time, and replace the crimp-on pins when there's any doubt. I must apologize for not posting my photos. My good camera is dead and I was never happy with my smart phone photos. I've also not had much spare time for forums lately.

Actually, I like your guess equally well, Gr5, and I find that my screws need tightening about half the time I encounter these symptoms. The other times, I must switch to my second cable (but I eventually need to repair both cables' open circuits or unreliable wire / connections). Unfortunately, carinac is still having trouble after tightening the screws, but I think I've just about eliminated the problem on my UMO so I'll take some photos today (sadly I'll need to use my smartphone, because my good digital camera is bust). I'll take photos of the metal crimp connectors and my crimp tool too and add a few more tips. There's no good way to correct a broken wire without cutting or pulling the wire apart at the break (slightly shortening the wire, but there's plenty of extra length available "upsteam" by re-doing the spiral wrap) and crimping on a new metal connector. The best way to check for a bad wire is to remove all three metal connectors from the white plastic shell and pull on them with moderate force. If you see the insulation stretch in one area, the wire is broken inside. If the wire breaks completely...well, it was probably broken anyway and you found the broken spot! Check all three wires too; you'll often find that more than one of them is bad. Give me a few hours to shovel some snow here in the wintry inland Pacific NW of the U.S. and I'll get some photos taken and posted for you.

I agree. My money's on the screws or the wires/connectors. I have many spare crimp-on connectors and have needed to cut the old ones off and crimp new connectors into place several times in the 5+ years of UMO use. Tiny tools (screwdriver or dental pick) are required to remove the crimp-on connectors from their shell. I use a crimping tool that is very close to the correct model for these connectors. Though it's not the best possible tool for the job, it's much better than needle nose pliers!. The wires often break inside the insulation and make unreliable contact. You can't see the break, because it's inside the insulation. The insulation should be crimped under the 2nd (larger) crimp ring of the connector and the wire NOT soldered. Solder stiffens the multi-strand wire and exacerbates the problem. Through trial and error I've figured out how to support the wires and minimize this problem. I do not route the wires through the black strain relief fingers, because then the wires just eventually break (inside the insulation) at the strain relief fingers. Rather, I go for a long unsupported "bridge" segment from fairly high up the Bowden tube down to the connector socket on the amplifirer board, with a "U" loop in the wire near the connector. I use a small dab of RTV silicone to tie the wire to the black stress relief fingers, but the RTV is flexible and allows some movement of the wire over a longer length of wire. You need to distribute wire movement and stress over as long a length of wire as possible or you will be plagued by frequent wire failures! Move the wires and move the head around. When I start to see tiny temperature reading fluctuations, I know that a wire/connector repair is in my near-term future, if tightening the wire screws doesn't fix it.

My StepStick 4988 drivers arrived from RepRap Champion today and they arrived with a big surprise, which might provide another option for those who would rather not rework their PCB in order to ground the 4988 ROSC pin to enable mixed decay mode on both rising and falling portions of the current waveform. The big surpise was the PBs are green instead of red! No--just kidding (though the boards ARE green instead of "StepStick Red," as pictured on the RepRap Champion website but the layout is obviously still a StepStick). No--the big surprise was that ROSC is already grounded, because the R4 component on the board is populated with a zero ohm resistor. Yippee! My aging eyes are spared from the board rework! Before I installed the new drivers, I measured my driver boards' Vrefs, as set by the Ultimaking, Ltd. Ultistepper manufacturer, and found them to be set from around 0.66 V to 0.68 V, which translates to 1.65 A to 1.70 A. According to the Pololu website, 1.7 A is the limit for the SY42STH38-1684A stepper motors so my UMO was set to run the steppers right up at maximum current. Even though the RepRap Champion website says their StepSticks can be adjusted for up to 2 A, I found that the highest I could set the Vref to be around 1.2 V to 1.3 V, depending on the board, before the wiper adjuster on the pot would max out (full clockwise) and drop Vref to zero volts, requiring me to back off the pot a touch. Given that my StepSticks use 0.1 ohm sense resistors (rather than the Ultistepper's 0.05 ohms), the maximum current setting for the RepRap Champion StepSticks is (in theory) 1.5 A to 1.625 A. I set them as high as possible and the slightly lower current (again, in theory) doesn't appear to affect the operation of my machine. I haven't printed anything yet but my stuttering Z stage is still gone too, which is what motivated this upgrade in the first place. RepRap Champion shipped the SpeedSticks incredibly fast (I ordered them Friday night and received them Monday from clear across the U.S.) so I still have a set of Pololu 8825 boards coming from elsewhere to try too. I'll let everyone here know if my occasional light zebra stripes change (hopefully for the better).

Yes. They may be your best bet, if you don't wish to modify your existing stepper motor cables with diodes yourself. You can also buy them here: https://thegr5store.com/store/index.php/umo/tl-smoother-removes-zebra-stripes.html On the other hand, different diodes or trying multiple diodes in series (depending on the diode's forward voltage drop) might yield different results, if you choose to modify your cables yourself. From https://www.reddit.com/r/3Dprinting/comments/5kful2/so_ive_recently_discovered_that_missing_steps/ The whole point of this board is to introduce a slight voltage drop into the wires, to compensate for using low-voltage steppers with a too high supply voltage. The S3G diodes have a slightly higher voltage drop, so if you were to use 1N5404's you'd need to replace each S3G diode with 2 1N5404's in series. They ripped this board/idea/concept off from Schrodinger Z who went with the 2 1N5404's route. In the end and for his DRV8825 controllers, Schrodinger Z used two 1N5404 in series for each direction, but he did try a single diode for each direction first: http://www.engineerination.com/2015/02/drv8825-missing-steps.html?m=1 1N5404 to 1N5408 diodes should all work about the same. I've studied the 4983, 4988, and 8825 datasheets and they all work slightly differently WRT mixed decay mode so your mileage may vary with various corrective actions. I don't know which controller the UM2 uses but, if it's a 4988, it would certainly be worth trying to ground the ROSC pin (either with a jumper to ground or by just bypassing an existing pull down resistor, if present, with a small "blue wire"), if you can handle the fine soldering job required. Unfortunately, if the controller is integrated onto the UM2 board, a lot more is at risk during such a mod than when hacking a cheap and replaceable "StepStick" or other Chinese controller board! If you're uncomfortable with it, then go with the diodes or commercial TL- Smoother diode modules. I have a set of Pololu 8825 drivers and a set of StepStick 4988 drivers ordered for my UM0. I plan to try diodes on both controllers and also a grounded ROSC configuration on the 4988s.

I finally decided to figure out why my UMO Z stepper has often stuttered when homing for a new print, requiring a power-off reset (an annoyance that I've lived with for over five years now). The short of my investigation (swapping drivers around) is my Z "Ultisputter" has always been defective! Looking on the bright side (an excuse to upgrade), I just ordered a new set of five Pololu DRV8825-based drivers. I figured they would be an improvement over my 4983-based drivers, but now I'm not so sure. After much online study, the zebra stripes seem to be most common with 8825 drivers and the diodes might be the only way to mitigate the problem. Also looking at the datasheets for the 4983, 4988, and 8225 chips, only the 4988 offers a 100% mixed decay mode (mixed decay is active for both rising and falling currents), which is what you're getting when you tie ROSC to ground. This configuration is reported here as a good one for the UMO/2/2+. I can't find a schematic for my old 4883-based Ultistepper and I'd like to know to know what value of ROSC was used. I'm not too keen on trying to probe the tiny board to find out (and I can't read the old and faded values on the old components either). Regardless, if the chosen ROSC is the usual 10k, my Ultisteppers are running the typically too short Toff timing. At least the 8225 offers a mixed decay mode, but it lacks the 100% (rising AND falling current) mixed decay mode that only the 4988 offers with a grounded ROSC configuration. You can download all the datasheets from Pololu. I'm an electrical engineer but have never bothered to study stopper motor control. Now I realize it's quite a critical but sometimes demanding design task! From the 4988 data sheet: For some applications where microstepping at very low speeds is necessary, the lack of back EMF in the winding causes the current to increase in the load quickly, resulting in missed steps. This is shown in Figure 2. By pulling the ROSC pin to ground, mixed decay is set to be active 100% of the time, for both rising and falling currents, and prevents missed steps as shown in Figure 3. So maybe I'd better order a set of 4988-based StepSticks too. They're cheap!

I have not tried any new ABS filaments in my somewhat customized and enhanced UMO for quite a long time. I have a large inventory of Ultimachine ABS stockpiled, and it was my favorite a couple years ago. I just got a couple of spools of MakerGeeks "High Performance PLA." I'm very impressed with it and prefer it to PolyMaker PolyMAX PLA, so far. The MakerGeeks high temperature PLA is not super expensive and it sure makes the hassles of printing ABS far less compelling! (BTW, I confirmed with their customer support that their "Dishwasher Safe PLA" is the exact same formulation as their "High Performance PLA.") I mostly make mechanical and structural parts and standard PLA is usually too brittle to be of much use to me. I mostly print nylon but large nylon parts are a hassle to print well too. I also purchased four spools of MakerGeeks PET-G and I plan to try it next next.

3.00 mm nominal is marginal in my UMO. I installed a slightly larger diameter Bowden tube for it, after I bought two cases of the original Taulman 618, which ran 3.00 mm nominal. Taulman now produces the industry standard 2.85.

I think in kisslicer the brim become smaller on the upper layers as well, i.e. fillet. Yes--steps are essential. Without the steps, you can end up with a net "minus" by inducing even more cooling stress into the part by making the brim thicker!

Kisslicer supports both a width and height input for brims. The best anti-warp / stress mitigation brim is not one layer. Rather, it is a a few or several layers with the width of the brim decreasing with height. Hence, from the build surface to the part, steps are built up the side of the part. The free (as in beer) version of Kisslicer is available for download so it's probably best to just inspect the results of various settings there: http://kisslicer.com/download.html For me, this is a critical feature for printing warp prone materials (nylon, and PC). Without it, I must design a similar structure into my solid models. I view ANY separation of the part from the build surface as a print failure. (It causes dimensional errors in the base as well as above the base, due to the resulting overextrusion from the lifting of the part.) BTW, Kisslicer dev, Jonathan, is resuming work on the program and plans to do a release soon, according to what he posted recently on the Kisslicer forum.

As long as we are requesting features, how about multiple layer stepped brims (like ages old Kisslicer)? They are critical for printing bed adhesion challenged materials like nylon. Simple brims are not adequate and rafts are messy and ugly. (Rafts ruin a nice smooth downskin). Kisslicer is the only software than supports this feature, AFAIK. If I don't use Kisslicer, I must painstaking design a stepped brim into my solid model or use "boolean glue" (Blender) to add it to the mesh, which often causes other problems (mesh errors or Blender complaints, etc.)

I agree with what many have said here. The Kinect is a low res camera and you'll probably upgrade from it pretty quickly. I used it with Reconstructme to scan the kids at my daughter's B-day party and I printed up busts for them all. I used a swivel chair without the back rest installed as a turntable. Even if Skanect works better for you, you'll still be very limited by the camera. I've been periodically trying to use Catch 123d for years but can never get anything useful from it.

I would probably not buy a more expensive printer for my business needs, because it would not "pencil-out" in terms of overall expense vs my time. I would either machine the parts myself (mostly prototypes) or send them out for manufacture (machining or rapid injection molding, or even silicone molded urethane sometimes). The UM (either UMO or UM2 or FFF/FDM in general) is right on the edge of being capable of making the kinds of industrial parts and prototypes that I require. When someone asks me about 3D printers, I tell them the experience is not like printing family vacation photos on an inkjet! I also tell them that it took me more time to learn how to use my UMO effectively than it took me to learn how to run a lathe or a milling machine. I can certainly understand your experiences with Slackware and your feelings about it too. Although there are more turn-key installation experiences nowadays (Ubuntu being the most popular), I have yet to install a Linux-based OS on any laptop without experiencing at least a few hassles--hassles that the typical Mac or Windows user would not be able to overcome. Desktop and server installations tend to go much more smoothly.

I have used the same since the early Slackware and Red Hat days. I don't find my productivity to be significantly different between the three; productivity is more a function of the application software design than the OS. I'm not a gamer but I understand your motivation for your choice there. I find that I can get Linux to do what I want more easily than the others when I need a customization, however. Because Linux is only an OS kernel rather than an OS, there are many OS options and styles that a user may choose (resulting in far more options and styles than with Apple and M$.) The source for nearly everything is available (proprietary HW drivers being the only typical exception), the scripting options are powerful, and everything you need is in the clear to see. There is no "Apple of 3D printers" and, in some aspects, I agree that it's a good thing! I've been very happy with my ability to customize Marlin. I still have my 3+ year-old UMO, which has a custom heated bed that supports a variety of interchangeable build surfaces, and a hot end that I also modified. I'm trying to find the time finish machining a Macor PEEK substitute to top 300 C, because I still prefer brass to stainless steel for the hot section. I ditched my short belts too. All of my mods were facilitated by the open design mindedness embodied in the UMO from its inception. None of the new Ultimaker machines are compelling to me, because they attempt to take a small step toward "Apple-ness" (superficially, I think) but sometimes deny me the ability to control the machine or work-around the common problems and difficulties associated with 3D printing. Some of the difficulties are associated with printing materials other than PLA, I know, and Ultimaker B.V. is still a very PLA-oriented company. I hardly ever print PLA, because I make strong structural and mechanical parts which will fail, if printed in conventional brittle PLA. It's been three years since I bought my Ultimaker (and the UMO was quite a "mature" design when I jumped-in) but still don't see any FFF/FDM printers on the market that can ooze molten plastic any better than my old UMO! FFF/FDM hardware and software development has actually been running at a very slow pace, compared to other emerging computer technologies. I think the biggest advances over the last three years have actually been made in the area of printing materials.

Deleted (double post).

Here's the workaround that I always use to slay that particular Cura annoyance. It doesn't work in "Quickprint" mode, but Quickprint poses only a minor annoyance (you can turn it off and it stays off): Use the the Layer Height box to take back control of your computer and command when Cura slices. Leave it set to "0" or fill the field with only a single decimal point. Cura will display a red background in the field and red is what you want to see (and feel ). When you are ready to slice, then type the number you want into the field. Be sure to go through the pains of "turning it off" again after each file is processed. No--it's not anywhere near as easy or efficient to use as a "start" button, like with every other slicer, but AFAIK it's all we've got in Cura. Cura has a couple of unique very useful features (like loading .ini settings from a gcode file) but autostart and Quickprint are not compelling features. Ditto for having to hit ctl+e to access the remainder of the settings menus, instead of having the "expert" setting selections consistent the menu tabs used for everything else (also only a very minor annoyance and of little consequence). There is no compelling reason to have Cura pointlessly chewing up CPU and memory resources while a user is getting setup to slice, or nearly anytime they make a setting change. When this "feature" appeared in Cura, I figured it wouldn't last very long after users found out what a bad idea it was, but I was wrong about this feature's tenure in Cura.

And what is the magnitude of the drain of resources needed to support Cura for the relatively small customer base running Linux (not that I'm against it)? Be honest now. I'd like your truthful assessment.

Cura is not a good reason for wasting my money and supporting a business that I despise (M$). I'm not a gamer. I don't need CPU and GPU-sucking eye candy and I explained my other reasons. My Dodge pickup truck works just as well as a new pickup truck for its utilitarian purposes (better in many ways). It is not at all crazy that I should wish to retain it too--especially given the cost to buy something new. I worked as a Senior Engineer at Intel for over 20 years. I know the computer industry's "tread mill" as well as anyone and it's sometimes just plain crazy! I'd rather suffer with FreeCAD and OpenSCAD than upgrade to another M$ OS offering that in no significant way improves my CAD capabilities over what I have now--or even aids Cura development, I suspect. I will counter your sage advice by telling you to boycott M$; it and its long parade of executive staff and former executive staff comprise some of the most evil members of both the computer industry and society in general. They are monopolistic economic fascists--pure and simple!

It's too bad that you let M$ dictate the development efforts of Cura and dropped support for XP. I suspect that a far greater number of people are still on XP than Linux in the world. As much as I hate M$ and love Linux, I think it will also be quite a long time before the installed world base of good ol' XP falls below the installed world base of Linux! Would it really be too hard to retain XP compatibility in Cura--even unofficially? I don't see much that's special in the way of Cura's presentation and look or functionality to justify some kind of new-fangled M$ dev kit or code library! I cannot afford to upgrade the software on my CAD station (and actually, my severe hatred for M$ keeps me from spending money on their new OSs that will not function any better in meeting my CAD needs than my existing system). I have an old 3.8 GHz Prescott running XP Pro x64 that features nearly full memory map access to all of its 8 GB of ECC memory. As an old "clock burner," it's actually still pretty fast for CAD applications, which even today are typically far from optimally threaded. If I "upgrade" my hardware or Windows to M$'s latest spyware-infested offerings (Snowden proved it with his outing of the NSA "Prism" program), I must also spend considerably more money for new CAD software licenses, in addition to the cost of the otherwise (other than Cura) completely unneeded OS. I don't like being forced to spin an industry "treadmill!" At least the latest release of Rhino5 still works with XP--even the 64-bit version, which is not officially supported. I tried to run Cura 15 in a Virtualbox Xubuntu VM, but it's too slow. I'm actually better off running it on my old dual Atom Ubuntu netbook that I use for general personal computing and travel.

Yes--most definitely! I've purchased parts from fbrc8 and it is an excellent vendor. I was very pleased when fbrc8 got a contract with UM Ltd. to be a support partner in the USA. Yes--fbrc8 would have been cheaper in the short term, but I don't need the short belts in the set and assuming that I keep my UMO for a long time or another direct drive UMO mod'er needs only a set of long belts, I think the B&B belts will work out to be more economical in the long run. They fit better than the original belts initially and have more room to stretch with any given tensioner. They fit so well that, hopefully my machine will not need any belt tensioners at all for a long while. Also, I probably could have continued to use the old belts with some other type of manual adjuster that could take up more of the slack, but clothespin springs are my favorites. They are "set and forget" and they feature a fairly wide adjustment range too. However, I can think of a couple of non-spring adjusters that might be able to take-up even more belt slack. On the other hand, the tooth spacing is continually increasing when belts age and stretch, so maybe a new tensioner would not have re-enabled 300+ mm/sec jumps for me, which helps with stringing and hairs in nylon prints.

I suspect that more than a few members here may have older UMs with aging stretched belts and might benefit from my experiences with belt replacements. I removed my short belt drive mechanisms and replaced them with direct drive over a year ago: https://www.thingiverse.com/thing:145544 You can checkout the "Made" link in the above page for my photos and notes. After installing the direct drive components, my UM became quieter and it required virtually no belt adjustments. (I re-bent my clothespin spring adjusters once.) I've tried several cool-looking printed manual tensioners, but I prefer the self-adjusting long belt clothespin spring tensioners, barely ever require any manual adjustment or maintenance. Almost immediately after assembling my new UMO over three years ago, I felt the supplied belts were a touch on the long side. After a short break-in period, the 3 mm screws and nuts in the belt "claws" of the slider blocks ran out of their extremely limited adjustment range. I suspect the screws and nuts were added by UM designers when they discovered that the claw did not result in sufficient tension of the 300 mm long belts and supplies of a slightly smaller belt size were problematic, due poor market availability. Regardless of the reasons for the 3 mm hardware, it is really a poor excuse for a belt tensioner! Thus I ran perfectly satisfactory, but kludgy-looking, clothespin springs for over three years. Recently, even bending the springs as much as possible, to maximize the tension of my aging belts, resulted in the springs rubbing gently on the UM's frame and still somewhat inadequate tension. (I could no longer "jump" reliably at 300 mm/sec!) I knew it was time to replace my long belts. Rather than buying 300 mm replacements, I decided to special order 298 mm belts from B&B Manufacturing in the U.S.: http://www.bbman.com/catalog/product/298MXL025G Unfortunately, BBMan charges some pretty hefty extra shipping and handling fees but also discounts quantity buy levels. Accordingly, I ordered 12 belts to more favorably amortize the $17.42 USD shipping and handling charges, which is an outrageous sum for an order quantity that fits in a medium size envelope! The total charges, delivered to my door about two months later, were $63.26 ($15.82 per long belt set), but now I probably have more than a lifetime supply of UMO belts! The 298MXL belts are the perfect size for use without any extra 3 mm claw tensioner hardware. (The wooden claw part alone results in perfect tension.) The 298 belts were slightly more difficult to install than the original 300s, and they were on the tight, but still acceptable side, in a belt pitch tone test before break-in (I think tone is the best way to check UMO belt tension), but they are perfect after a couple of prints worth of break-in time: (I don't know why UM doesn't emphasize the above method, because any belt tension evaluation that is based on belt deflection depends on providing a known and standard pull force on the belt!) I expect to get maximum service life from these belts, because no extra tensioner is needed initially, as it was with my original UMO 300 mm belts. I'll probably add my clothespin springs again, when necessary, for set and forget convenience, but the original 3 mm claw hardware could also be used to extend my 298 mm belt service life beyond what it can accomplish in tensioning the original belts too.

Or get an Ultimaker! Seriously, this is one of the best forums, I believe, and the open source availability and usability (UM has really improved it, of late) is a very compelling selling feature of the Ultimaker. Though any 3D printer comes with a road full of bumps, I have never regretted buying an Ultimaker almost three years ago.

Have you found more than hearsay that it's "nasty stuff" at the temperatures we print? I'm talking about empirical testing! I'm not saying that it isn't nasty, but how convenient that a reasonably priced and excellent performing source of nylon would be deemed inappropriate and unsafe to use by those who sell printer filament! As with many activities, I recommend playing it safe by not printing where one breathes (that goes for any plastic--even PLA). Ink jet printers have been found to out-gas unhealthy fumes in an Australian study and they are found in all office buildings and many homes too!